Welcome to MyoSuite’s documentation!

MyoSuite is a collection of musculoskeletal environments and tasks simulated with the MuJoCo physics engine. It serves researchers and practitioners across biomechanics, neuroscience, machine learning, sports medicine, and physical rehabilitation.

GitHub | Paper (arXiv) | Slack

Note

This project is under active development.

What’s new in MyoSuite 3

MyoSuite 3 brings the whole suite to fast, scalable training while keeping the simple interface you know:

• One task, several backends. The same env_id runs on your CPU through the standard Gymnasium interface (to explore, debug and replay policies, or to train with libraries such as Stable-Baselines3), on mjlab for massively parallel GPU training, and on an experimental MJX (JAX) path. See Machine Learning Quick Start.

• MuscleMimic support. Run, evaluate and train full-body and bimanual MuscleMimic policies, with ready-to-use checkpoints and motion datasets.

• The complete MyoChallenge suite as Gymnasium environments: Baoding, Bimanual, Chase Tag, Die Reorient, OSL Run, Relocate, Soccer and Table Tennis.

• Much faster learning. Train with thousands of environments in parallel on a single GPU, then replay the policy on the CPU.

• New tasks in a few lines. Describe a task with a compact spec and reuse the shared observation, reward and model-building blocks instead of writing an environment class.

• Some existing environments changed. Observations are no longer clipped and are read after a fresh forward step (55 env ids), the motorFinger* envs have 4x stronger motors, the Random finger-reach tasks now sample only targets the fingertip can reach, and several reset and seed behaviours were corrected. Policies trained with MyoSuite 2.x or earlier snapshots may need retraining; the repository’s CHANGELOG.md lists the changes.

• Ready to use. Default trained policies with evaluation videos (see RL baselines), plus updated tutorials from the first rollout to GPU training and MuscleMimic (Tutorials). The repository’s CHANGELOG.md lists everything that changed since v2.12.

Choose your path

I am a…	Start here
Biomechanist	Biomechanics Quick Start — kinematics, muscle forces, inverse dynamics, OpenSim
Neuroscientist	Neuroscience Quick Start — proprioception, reflex controllers, fatigue
ML / RL Researcher	Machine Learning Quick Start — Gymnasium API, SB3, mjlab GPU training
Sports / Rehab Clinician	Sports Medicine & Rehabilitation Quick Start — pathological conditions, clinical metrics

Quick Start by Audience

• Biomechanics Quick Start
• Neuroscience Quick Start
• Machine Learning Quick Start
• Sports Medicine & Rehabilitation Quick Start

Installation & Tutorials

• Installation
• Tutorials

Reference

• Architecture
• Environment Reference
• ModelBuilder (experimental)
• Backend parity
• Fatigue Model Validation

Advanced Features

• Models and Tasks
• myoFinger
• myoElbow
• myoHand
• myoLeg
• myoArm
• myoTorso
• myoArmNoHandMuscles
• Finger Joint Pose
• Finger Tip Reach
• Elbow Joint Pose
• Hand Joints Pose
• Hand Tips Reach
• Hand Key Turn
• Hand Object Hold
• Hand Pen Twirl
• Arm Reach
• Hand Multiobject Reorientation
• Hand Die Rotation (MyoChallenge 2022)
• Hand Baoding Balls (MyoChallenge 2022)
• Leg Walk

Projects with MyoSuite

• Projects with Myosuite
• RL baselines
• MyoChallenge-2024 Documentations
• MyoChallenge-2025 Documentations

API Reference

• API Reference
  • Gymnasium Wrappers
  • Muscle Action Term
  • Muscle-Command Stages
  • Policy Utilities
  • Physics & Biomechanics
  • Quaternion Math Utilities
  • Tensor Utilities
  • Renderer

References

• Publications
• References
• API Reference

How to cite

If you use MyoSuite, please cite the MyoSuite paper:

@inproceedings{Caggiano2022MyoSuite,
   title     = {{MyoSuite} -- A contact-rich simulation suite for musculoskeletal motor control},
   author    = {Caggiano, Vittorio and Wang, Huawei and Durandau, Guillaume and Sartori, Massimo and Kumar, Vikash},
   booktitle = {Learning for Dynamics and Control (L4DC)},
   year      = {2022},
   doi       = {10.48550/ARXIV.2205.13600},
   url       = {https://arxiv.org/abs/2205.13600},
}

For MyoSuite 3, cite it as software:

@misc{MyoSuite2026,
   title        = {{MyoSuite} 3.0 -- A multimodal platform for efficient and scalable musculoskeletal motor control},
   author       = {Caggiano, Vittorio and Hodossy, Balint and Fischer, Florian and Wang, Cheryl and {MyoSuite Team}},
   year         = {2026},
   howpublished = {\url{https://github.com/myohub/myosuite}},
}

For the full-body MuscleMimic models, checkpoints, retargeted motions and tutorials 5.1-5.5, also cite:

@article{Li2026MuscleMimic,
   title   = {Towards Embodied AI with {MuscleMimic}: Unlocking full-body musculoskeletal motor learning at scale},
   author  = {Li, Chengkun and Wang, Cheryl and Ziliotto, Bianca and Simos, Merkourios and Kovecses, Jozsef and Durandau, Guillaume and Mathis, Alexander},
   journal = {arXiv preprint arXiv:2603.25544},
   year    = {2026},
}

The motion data come from AMASS and its KIT subset, retargeted with GMR; the datasets are for non-commercial research and ask you to cite AMASS and MuscleMimic. See References for these and all other sources (models, muscle fatigue, controllers, motion data).